This skill should be used when the user asks to "check my portfolio risk", "scan my holdings", "are my tokens safe", "portfolio guard", "持仓风控", "check if any of my tokens are risky", or "monitor my assets".
A comprehensive holdings risk monitor that scans every token in the user's wallet across chains, classifies risk level, and optionally executes disposal of dangerous positions.
Collect a complete picture of what the user holds before any scanning begins.
wallet-status to identify the currently connected wallet address and active chain.wallet-balance on the current chain to retrieve all token balances with their USD values.wallet-balance again on the other chain(s) to pick up cross-chain holdings.Do not proceed if wallet-status returns no connected wallet. Ask the user to connect a wallet first.
Not every token needs a deep scan. Apply the following filter:
Skip (mark 🟢 Safe immediately, no further scanning needed):
Queue for deep scan — all other tokens, including:
List the scan queue clearly so the user knows which tokens are being analyzed.
Run scans in parallel where the underlying APIs allow concurrency. For each queued token, perform all four sub-scans:
security-token-scanCall security-token-scan with the token's contract address and chain. Extract and evaluate:
Record raw findings for each flag.
Query current market data for the token (price feed, DEX aggregator, or on-chain liquidity pool):
Query recent on-chain transaction activity:
Analyze the 7-day price history:
After all scan data is collected, classify each scanned token into one of three tiers:
Assign 🔴 if ANY of the following are true:
security-token-scan flags: honeypot detectedsecurity-token-scan flags: mint backdoor active (deployer can mint)security-token-scan or external feed: token explicitly flagged as scam / rug pullAssign 🟡 if ANY of the following are true (and no 🔴 criteria met):
security-token-scan flags: ownership not renounced AND contract is upgradeable (proxy)security-token-scan flags: transfer tax >5%Assign 🟢 if:
security-token-scan returns clean (no honeypot, no mint backdoor, no high tax)Present a structured risk report. Always sort sections by severity (🔴 first, then 🟡, then 🟢).
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
SentryX PORTFOLIO GUARD — RISK REPORT
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
PORTFOLIO OVERVIEW
Total Value: $XX,XXX
Tokens Scanned: X of Y holdings
Chains: Ethereum, Solana (example)
Risk Summary:
🔴 Immediate Action: X token(s)
🟡 Close Monitoring: X token(s)
🟢 Healthy / Safe: X token(s)
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
🔴 IMMEDIATE ACTION REQUIRED
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
[TOKEN SYMBOL] — [Chain]
Value: $X,XXX (X.X% of portfolio)
Contract: 0x...
Findings:
• HONEYPOT DETECTED — sells are blocked by contract
• Deployer wallet sold 80% of holdings in past 3 days
Recommendation: EXIT IMMEDIATELY if possible. Note: honeypot
means selling may not be possible. See Step 6.
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
🟡 CLOSE MONITORING
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
[TOKEN SYMBOL] — [Chain]
Value: $XXX (X.X% of portfolio)
Contract: 0x...
Findings:
• Liquidity dropped 55% in past 7 days
• Top 10 holders control 81% of supply
• Transfer tax: 8% (high)
Recommendation: Monitor closely. Consider reducing position.
Set a stop-loss at current liquidity levels.
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
🟢 HEALTHY / SAFE
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
ETH — Ethereum $X,XXX (XX.X%) Native token
SOL — Solana $X,XXX (XX.X%) Native token
USDC — Ethereum $X,XXX (XX.X%) Stablecoin
[TOKEN] — [Chain] $XXX (X.X%) No risk signals detected
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
ACTIONS AVAILABLE
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
• "sell all risky" / "卖出高危" — Attempt to sell all 🔴 tokens
• "sell [TOKEN]" — Sell a specific token
• "ignore all" / "忽略" — Acknowledge and dismiss
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
After presenting the report, wait for the user's response. Handle the following commands:
For each 🔴 token in order of severity:
security-tx-scan to simulate the sell transaction before execution. This is mandatory. Do not skip.security-tx-scan or security-token-scan indicates the token is a honeypot (sell transaction will revert / no sell path exists):
Repeat for all 🔴 tokens. After all are processed, summarize: how many sold successfully, how many were honeypots or failed, total USD value recovered.
Same flow as above but for a single named token. The token may be 🔴, 🟡, or even 🟢 — respect the user's explicit instruction. Run simulation, check honeypot, confirm, execute.
Acknowledge the user's decision without judgment. Summarize the findings briefly and end the skill run. Remind the user they can run portfolio-guard again at any time.
Never fabricate data. If a market data API returns no result for a token, report it as "data unavailable" and note the uncertainty. Do not invent prices, liquidity figures, or holder concentrations.
Honeypot = explicit disclosure. If a token is flagged as a honeypot, always tell the user clearly and explicitly. Never silently skip it or imply that selling might work when it won't.
Always simulate before selling. security-tx-scan simulation is mandatory before any sell execution. This protects the user from unexpected reverts, sandwich attacks, and extreme slippage.
Sort by severity. Always present 🔴 tokens before 🟡 before 🟢 in the report. Urgency must be visually clear.
Never execute without confirmation. Always show the user a preview of the transaction (amount, expected output, gas, slippage) and require explicit confirmation before submitting on-chain.
Cross-chain awareness. If the user's 🔴 token is on a different chain than the currently active chain, note that a chain switch may be required before selling and guide the user through it.
LP tokens. If the user holds LP (liquidity provider) tokens, treat them as a special case. Flag that removing LP requires calling the liquidity removal function, not a standard token swap. Warn if the underlying pool has low liquidity (risk of impermanent loss on exit).
When executing sells in Step 6, apply the full safe-trade logic from the safe-trade skill:
security-tx-scanDelegate to safe-trade patterns for execution rather than reimplementing sell logic here.